The New England Journal of Medicine

The New England Journal of Medicine Copyright, 1998, by the Massachusetts Medical Society VOLUME 339 A UGUST 20, 1998 NUMBER 8 EFFECT OF BETA-BLOCKADE ON MORTALITY AMONG HIGH-RISK AND LOW-RISK PATIENTS AFTER MYOCARDIAL INFARCTION STEPHEN S. GOTTLIEB, M.D., ROBERT J. MCCARTER, SC.D., AND ROBERT A. VOGEL, M.D. ABSTRACT Background Long-term administration of betaadrenergic blockers to patients after myocardial infarction improves survival. However, physicians are reluctant to administer beta-blockers to many patients, such as older patients and those with chronic pulmonary disease, left ventricular dysfunction, or non Q-wave myocardial infarction. Methods The medical records of 201,752 patients with myocardial infarction were abstracted by the Cooperative Cardiovascular Project, which was sponsored by the Health Care Financing Administration. Using a Cox proportional-hazards model that accounted for multiple factors that might influence survival, we compared mortality among patients treated with beta-blockers with mortality among untreated patients during the two years after myocardial infarction. Results A total of 34 percent of the patients received beta-blockers. The percentage was lower among the very elderly, blacks, and patients with the lowest ejection fractions, heart failure, chronic obstructive pulmonary disease, elevated serum creatinine concentrations, or type 1 diabetes mellitus. Nevertheless, mortality was lower in every subgroup of patients treated with beta-blockade than in untreated patients. In patients with myocardial infarction and no other complications, treatment with beta-blockers was associated with a 40 percent reduction in mortality. Mortality was also reduced by 40 percent in patients with non Q-wave infarction and those with chronic obstructive pulmonary disease. Blacks, patients 80 years old or older, and those with a left ventricular ejection fraction below 20 percent, serum creatinine concentration greater than 1.4 mg per deciliter (124 µmol per liter), or diabetes mellitus had a lower percentage reduction in mortality. Given, however, the higher mortality rates in these subgroups, the absolute reduction in mortality was similar to or greater than that among patients with no specific risk factors. Conclusions After myocardial infarction, patients with conditions that are often considered contraindications to beta-blockade (such as heart failure, pulmonary disease, and older age) and those with nontransmural infarction benefit from beta-blocker therapy. (N Engl J Med 1998;339:489-97.) 1998, Massachusetts Medical Society. SEVERAL large trials have demonstrated that the long-term administration of beta-adrenergic blockers to patients after myocardial infarction improves survival. 1-3 Although the percentage of such patients receiving beta-blockade has recently increased, physicians still reportedly prescribe beta-blockers for less than one third, 4 and cardiologists for less than half, 5 of patients with myocardial infarction. Older age, impaired left ventricular function, transient heart failure, and the use of diuretic drugs predict lack of use of these drugs. 6 This suggests that physicians are concerned that betablockers may be contraindicated in a substantial proportion of patients. The Cooperative Cardiovascular Project, a program to evaluate the care of Medicare patients with a diagnosis of myocardial infarction, provides the opportunity to evaluate the relation between treatment and outcome in unselected patients, most of whom are over 65 years of age. With more than 200,000 charts abstracted, this data base gives investigators the power to evaluate the effect of interventions in situations rarely studied before. Moreover, this data base provides information on unselected patients with conditions that often keep them out of randomized, controlled trials. We sought to use the Cooperative Cardiovascular Project data to determine which patients benefit from the use of beta-blockers. We compared the mortality rates of patients in high- and low-risk subgroups who were treated with beta-blockers with the rates among those not given those drugs. Correcting to the extent possible for differences between the groups, we evaluated the effects of beta-blockers in patients with presumed contraindications to their use, such as old- From the Department of Medicine and the Department of Epidemiology and Preventive Medicine, University of Maryland School of Medicine, Baltimore. Address reprint requests to Dr. Gottlieb at the Division of Cardiology, University of Maryland Medical Systems, 22 S. Greene St., Baltimore, MD 21201. Volume 339 Number 8 489

er age, low ejection fraction, chronic obstructive pulmonary disease, diabetes mellitus, low blood pressure, and low heart rate. In addition, the data base provided the opportunity to look at beta-blockade in the setting of contemporary medical practice. METHODS The Cooperative Cardiovascular Project data base includes all acute care hospital claims for patients with a principal diagnosis of acute myocardial infarction that were submitted to the Health Care Financing Administration for Medicare payment. A diagnosis of acute myocardial infarction was defined according to the coding system of the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM), as code 410, with those with a fifth digit of 2 (for episode of care), those reported by nonacute care hospitals, and those with missing medical-record numbers excluded. The national sample consisted of all discharges involving cases of acute myocardial infarction during an eightmonth period, with data from the various states collected so that most discharges fell between February 1994 and July 1995. The total number of discharges in the national sample was 201,752. The abstraction was performed by trained medical abstractors using uniform criteria to define each data item. They had access to all original hospital records. Extensive quality-control measures were applied to validate the reliability and accuracy of the abstracts. Finally, extensive checks were run against the computerized abstracts to ensure internal consistency and to enforce common data standards. 7 The data were nearly complete. Serum creatinine concentrations were missing for 9575 patients, and most other data were missing for even fewer patients. The ejection fraction was obtained or documented in 134,238 patients. Patients for whom the ejection fraction was missing were analyzed as a separate subgroup. To avoid eliminating the few patients who had other missing data from the analyses, we combined patients with missing data into the largest group for all variables other than the one being analyzed. Whenever possible, the data reflect the values obtained on admission to the hospital. The Health Care Financing Administration s survival data were based on Social Security records and were complete, with accuracy of 98 to 99 percent, at 60 days. In this study, the Cooperative Cardiovascular Project data were used to address the question of who benefits from beta-blocker therapy. Initially, the patients prescribed and not prescribed betablocker therapy were compared with regard to a number of demographic and clinical variables. Patients for whom any beta-blocker at any dose had been prescribed at the time of discharge were classified as having received a beta-blocker. The means and standard deviations of continuous variables were compared. Both continuous (in categories) and discrete variables were compared with contingency-table analyses. Comparison of mortality rates between patients receiving and those not receiving beta-blocker therapy was based on a time-toevent analysis. The Cox proportional-hazards model was used for this purpose, because it permitted estimated survival time to be controlled for differences in follow-up time and differences in covariates that might affect survival, including both confounding variables and effect modifiers. The proportional-hazards assumption was initially checked with Kaplan Meier analyses and was found to be adequately met. Only 8464 of the total of 201,752 patients were not followed for 24 months. Because of the size of the data set, emphasis was placed on the magnitude rather than the statistical significance of differences between groups. We recognized that very small differences could (and did) achieve statistical significance. RESULTS Thirty-four percent of the patients received betablockers (Table 1). The percentage was lower among the very elderly, blacks, and those patients considered the sickest according to their Acute Physiology and Chronic Health Evaluation (APACHE II) 8 and Killip 9 scores. Similarly, patients with the lowest ejection fractions, heart failure, chronic obstructive pulmonary disease, elevated serum creatinine concentrations, or type 1 diabetes mellitus were less likely to be treated with beta-blockade. Conversely, patients who had been given more aggressive treatment, including thrombolysis and angioplasty, but not bypass surgery, were more likely to be receiving beta-blockers. As a result of these prescribing practices, patients receiving beta-blockers tended to have fewer risk factors for mortality than those not receiving beta-blockers. They were younger (mean [±SD] age, 73.3±8.8 vs. 74.5±9.0 years) and had lower (i.e., more favorable) APACHE II scores (7.9±3.4 vs. 9.0±4.4) and lower Killip scores. Their hospital stay was shorter (7.2±5.1 vs. 7.8±7.5 days), and their physiologic condition, as indicated by the serum creatinine concentration (1.3±1.0 vs. 1.4±1.1 mg per deciliter [115±88 vs. 124±97 µmol per liter]) and left ventricular ejection fraction (47±13 percent vs. 44±14 percent), was better. They had fewer coexisting diseases, including chronic obstructive pulmonary disease, diabetes, and congestive heart failure. Finally, they were more likely to be given thrombolytic and reperfusion therapy. The only exceptions to the favorable profile among those prescribed beta-blockers were the slightly higher proportions of patients with hypertension and of men in this group. Because of the better risk-factor profile among those treated with beta-blockers, it was necessary to hold each risk factor constant when comparing mortality among those treated and those not treated with beta-blockers. Each of the subsequent analyses took direct account of or held constant the following factors: age; race or ethnic group; sex; APACHE II score; length of hospital stay; Killip score; physiologic condition at admission, as reflected by the systolic blood pressure, pulse rate, and serum creatinine concentration; left ventricular ejection fraction; presence or absence of coexisting diseases, including diabetes, congestive heart failure, chronic obstructive pulmonary disease, and asthma; type of myocardial infarction (anterior, inferior, or non Q-wave); history of myocardial infarction, revascularization, or heart failure; and treatment, including aspirin, angiotensin-converting enzyme inhibitors, calcium-channel blockers, heparin, thrombolytic agents, coronary angioplasty, and bypass surgery. Table 2 shows the 24-month adjusted risks and relative risks of death among patients prescribed beta-blockers and those not prescribed the drugs. After demographic and physiologic measures and coexisting diseases were accounted for, patients treated with beta-blockade had substantially lower 490 August 20, 1998

EFFECT OF BETA-BLOCKADE ON MORTALITY IN HIGH-RISK AND LOW-RISK PATIENTS AFTER MYOCARDIAL INFARCTION TABLE 1. CHARACTERISTICS OF PATIENTS WHO RECEIVED OR DID NOT RECEIVE S AT HOSPITAL DISCHARGE.* CHARACTERISTIC NO. OF PATIENTS NO CHARACTERISTIC NO. OF PATIENTS NO no. (%) no. (%) Total Age <75 yr 75 84 yr»85 yr Sex Male Female Race or ethnic group White Black Hispanic Other APACHE II score <8 8 10.9 11 12.9»13 Highest Killip score 1 or 2 3 4 Left ventricular ejection fraction <20% 20 29% 30 39% 40 49%»50% Missing data Systolic blood pressure <100 mm Hg 100 119 mm Hg 120 139 mm Hg»140 mm Hg 201,752 106,496 71,426 23,830 109,193 92,559 180,320 13,681 5,832 1,919 95,698 61,402 18,023 26,629 113,151 82,346 6,255 2,812 13,284 23,968 33,860 60,314 67,514 10,457 29,910 48,950 111,926 69,153 (34.3) 39,312 (36.9) 23,467 (32.9) 6,374 (26.7) 38,236 (35.0) 30,917 (33.4) 62,531 (34.7) 4,394 (32.1) 1,660 (28.5) 568 (29.6) 37,517 (39.2) 21,150 (34.4) 5,074 (28.2) 5,412 (20.3) 46,547 (41.1) 21,336 (25.9) 1,270 (20.3) 412 (14.7) 2,975 (22.4) 7,595 (31.7) 13,350 (39.4) 24,787 (41.1) 20,034 (29.7) 2,679 (25.6) 9,512 (31.8) 16,838 (34.4) 40,000 (35.7) 132,599 (65.7) 67,184 (63.1) 47,959 (67.1) 17,456 (73.3) 70,957 (65.0) 61,642 (66.6) 117,789 (65.3) 9,287 (67.9) 4,172 (71.5) 1,351 (70.4) 58,181 (60.8) 40,252 (65.6) 12,949 (71.8) 21,217 (79.7) 66,604 (58.9) 61,010 (74.1) 4,985 (79.7) 2,400 (85.3) 10,309 (77.6) 16,373 (68.3) 20,510 (60.6) 35,527 (58.9) 47,480 (70.3) 7,778 (74.4) 20,398 (68.2) 32,112 (65.6) 71,926 (64.3) Serum creatinine <1.2 mg/dl 1.2 1.4 mg/dl 1.5 1.9 mg/dl»2.0 mg/dl Heart rate <60/min 60 79/min 80 99/min 100 119/min»120/min Myocardial infarction Q-wave Non Q-wave COPD Asthma Diabetes mellitus Receiving insulin Not receiving insulin No diabetes Prior CHF Prior myocardial infarction Prior CABG Prior PTCA Treatment during current hospitalization CABG PTCA Thrombolytic agents ACE inhibitors Calcium-channel blockers Aspirin 119,557 27,464 27,026 18,129 17,788 67,651 65,343 31,782 18,756 116,734 85,018 41,814 3,819 21,578 37,867 139,150 39,838 63,142 26,872 16,532 18,947 31,576 29,642 60,382 65,609 166,650 44,094 (36.9) 8,979 (32.7) 7,866 (29.1) 4,691 (25.9) 6,397 (36.0) 26,970 (39.9) 22,421 (34.3) 8,898 (28.0) 4,350 (23.2) 40,715 (34.9) 28,438 (33.4) 9,228 (22.1) 676 (17.7) 6,115 (28.3) 12,521 (33.1) 49,477 (35.6) 8,633 (21.7) 21,046 (33.3) 9,735 (36.2) 6,216 (37.6) 5,946 (31.4) 14,365 (45.5) 11,755 (39.7) 19,619 (32.5) 20,206 (30.8) 61,953 (37.2) 75,463 (63.1) 18,485 (67.3) 19,160 (70.9) 13,438 (74.1) 11,391 (64.0) 40,681 (60.1) 42,922 (65.7) 22,884 (72.0) 14,406 (76.8) 76,019 (65.1) 56,580 (66.6) 32,586 (77.9) 3,143 (82.3) 15,463 (71.7) 25,346 (66.9) 89,673 (64.4) 31,205 (78.3) 42,096 (66.7) 17,137 (63.8) 10,316 (62.4) 13,001 (68.6) 17,211 (54.5) 17,887 (60.3) 40,763 (67.5) 45,403 (69.2) 104,697 (62.8) *APACHE denotes Acute Physiology and Chronic Health Evaluation, COPD chronic obstructive pulmonary disease, CHF congestive heart failure, CABG coronary-artery bypass graft, PTCA percutaneous transluminal coronary angioplasty, and ACE angiotensin-converting enzyme. To convert values for creatinine to micromoles per liter, multiply by 88.4. Lower scores denote less severe illness. mortality in every subgroup. Among patients with myocardial infarction and no other complications, treatment with beta-blockers was associated with a 40 percent reduction in mortality during the followup period. Mortality was also reduced by 40 percent among patients with non Q-wave infarction (Fig. 1) and patients with chronic obstructive pulmonary disease (Fig. 2). However, the reduction in mortality among those treated with beta-blockers was smaller among black patients (28 percent). Older patients (Fig. 3) and those with a left ventricular ejection fraction below 20 percent (Fig. 4), a serum creatinine concentration higher than 1.4 mg per deciliter, or diabetes mellitus also had a smaller percentage reduction in mortality. However, because of the higher mortality rates in these subgroups, the use of beta-blockers actually resulted in an absolute reduction in mortality similar to or greater than that in patients with no specific risk factors. DISCUSSION Patients who received beta-blockers had a 40 percent lower mortality rate than those who did not have beta-blockers prescribed at the time of discharge from the hospital. This reduction was even greater than the 22 percent reduction found in a meta-analysis of 23 trials. 10 Furthermore, all subgroups of patients had similar reductions in mortality in the present study. Nonetheless, only 34 percent of patients were discharged with a prescription for beta-blockers. Lowrisk patients were more likely to be treated with beta-blockade, but nevertheless only 41 percent of patients with normal ejection fractions received betablockers. This finding is consistent with the results of previous studies that have shown low rates of use of beta blockers for young, ideal candidates. 5 The most severely ill patients were even less likely to receive beta-blockers. If one assumes a mean reduction of 9.5 deaths per 100 patients with myocardial Volume 339 Number 8 491

TABLE 2. ADJUSTED RISK AND RELATIVE RISK OF DEATH AMONG PATIENTS WHO RECEIVED OR DID NOT RECEIVE S AT HOSPITAL DISCHARGE.* CHARACTERISTIC RECEIVING RISK OF DEATH AT 2 YR NOT RECEIVING DIFFERENCE IN RISK RELATIVE RISK (95% CI) percent Patient without complications 14.4 23.9 9.5 Age (yr) <70 yr 70 79 yr»80 yr 11.3 15.3 22.6 18.7 24.0 33.1 7.4 8.7 10.5 0.64 (0.58 0.70) 0.68 (0.63 0.75) Black race 16.5 23.0 6.4 0.72 (0.66 0.79) Prior COPD 16.8 27.8 11.1 Asthma 11.9 19.7 7.8 Diabetes mellitus 17.0 26.6 9.6 0.64 (0.60 0.69) Q-wave MI 14.2 23.6 9.4 Non Q-wave MI 14.4 23.9 9.5 Prior CHF 17.4 28.9 11.5 Prior MI 16.8 25.1 8.4 0.67 (0.62 0.72) Systolic blood pressure <100 mm Hg 100 139 mm Hg»140 mm Hg Ejection fraction <20% 20 49%»50% Missing data Serum creatinine <0.8 mg/dl 0.8 1.4 mg/dl >1.4 mg/dl Heart rate <70/min 70 99/min»100/min Treatment during current hospitalization CABG PTCA Thrombolytic agents Calcium-channel blockers ACE inhibitors Aspirin 16.9 10.4 9.8 23.5 15.3 11.6 12.3 12.9 13.9 19.4 13.1 14.9 16.9 6.1 9.2 11.8 16.4 14.4 13.8 28.1 17.2 14.8 34.5 25.4 19.3 20.4 21.4 23.1 29.9 21.7 24.8 26.2 10.2 15.2 19.6 23.6 23.9 22.9 11.2 6.8 5.0 11.0 10.1 7.7 8.1 8.5 9.2 10.5 8.6 9.9 9.3 4.0 6.0 7.8 7.1 9.5 9.1 0.66 (0.61 0.71) 0.68 (0.58 0.80) 0.65 (0.61 0.69) 0.65 (0.60 0.69) 0.70 (0.65 0.75) *CI denotes confidence interval, COPD chronic obstructive pulmonary disease, MI myocardial infarction, CHF congestive heart failure, CABG coronary-artery bypass graft, PTCA percutaneous transluminal coronary angioplasty, and ACE angiotensin-converting enzyme. To convert values for creatinine to micromoles per liter, multiply by 88.4. Because of rounding, the difference in risk does not always equal the difference between the risks shown for patients receiving a beta-blocker and those not receiving a beta-blocker. For patients without complications, all risk factors were set at the Cooperative Cardiovascular Project population means, with none of the factors that alter the effect of beta-blockers. infarction in 24 months, more than 19,000 patients could have been kept alive had beta-blockers been more widely prescribed. Non Q-Wave Infarction The relative risk of death in patients with non Q-wave infarction who were prescribed beta-blockers was markedly decreased. This is important, because the use of beta-blockers in such patients has been questioned since the b-blocker Heart Attack Trial found no improvement in survival among patients with nontransmural infarction who were given beta-blockers. 11 However, this study was a subgroup analysis performed at a time when the treatment of non Q-wave infarction was substantially different from present practice. Indeed, the American College of Cardiology and American Heart Association joint guidelines for the management of acute myocardial infarction classify the evidence for the use of betablockade in this subgroup as conflicting and possibly unfavorable. 12 The overwhelming evidence from our study, however, suggests that beta-blockade should 492 August 20, 1998

EFFECT OF BETA-BLOCKADE ON MORTALITY IN HIGH-RISK AND LOW-RISK PATIENTS AFTER MYOCARDIAL INFARCTION 1.0 Proportion Surviving 0.8 0.6 0.4 0.2 No beta-blocker, non Q-wave MI No beta-blocker, Q-wave MI Beta-blocker, non Q-wave MI Beta-blocker, Q-wave MI 0.0 0 6 12 18 24 Months NO. ALIVE 201,752 173,906 162,745 152,298 138,077 Figure 1. Adjusted Probability of Survival among Patients with Q-Wave and Non Q-Wave Myocardial Infarction (MI) Who Received or Did Not Receive Beta-Blockers. The patients with Q-wave infarction and those with non Q-wave infarction had similar benefit with beta-blockade. 1.0 Proportion Surviving 0.8 0.6 0.4 0.2 No beta-blocker, no COPD No beta-blocker, COPD Beta-blocker, no COPD Beta-blocker, COPD 0.0 0 6 12 18 24 Months NO. ALIVE 201,752 173,906 162,745 152,298 138,077 Figure 2. Adjusted Probability of Survival among Patients with or without a History of Chronic Obstructive Pulmonary Disease (COPD) Who Received or Did Not Receive Beta-Blockers. Patients with chronic obstructive pulmonary disease had a larger absolute benefit with beta-blockade. be part of the treatment of patients with non Q-wave infarction. Older Age The present study demonstrates that older patients have a greater absolute benefit when treated with beta-blockade. Indeed, even among those 80 years of age or older, mortality was 32 percent lower when beta-blockers were prescribed. It is thus unfortunate that the elderly often receive less treatment after a myocardial infarction than younger patients. 13 The present study found that only 27 percent of those over 84 years of age received beta-blockers. In a previous study of elderly patients admitted to a nursing home with evidence of a prior myocardial infarction, only 8 percent were receiving beta-blockade. Interestingly, in that study 37 percent of the patients were receiving calcium-channel blockers, which have not been shown to improve survival and may actually increase mortality. 14 There are few studies of older patients, which might explain the reluctance on the part of many physicians to use these agents. Previous studies, even of elderly patients, have generally limited analysis to patients less than 69 or 75 years of age. 15 Nevertheless, the results of our study are consistent with those Volume 339 Number 8 493

1.0 Proportion Surviving 0.8 0.6 0.4 0.2 No beta-blocker, <70 yr No beta-blocker, 70 79 yr No beta-blocker,»80 yr Beta-blocker, <70 yr Beta-blocker, 70 79 yr Beta-blocker,»80 yr 0.0 0 6 12 18 24 Months NO. ALIVE 201,752 173,906 162,745 152,298 138,077 Figure 3. Adjusted Probability of Survival According to Age among Patients Who Received or Did Not Receive Beta-Blockers. The oldest patients had a smaller relative benefit (32 percent reduction in the risk of death) but a larger absolute benefit than younger patients. The curve for the patients not receiving beta-blockers who were 70 to 79 years old overlaps that for the patients receiving beta-blockers who were»80 years old. 1.0 Proportion Surviving 0.8 0.6 0.4 0.2 No beta-blocker,»50% No beta-blocker, 20 49% No beta-blocker, <20% Beta-blocker,»50% Beta-blocker, 20 49% Beta-blocker, <20% 0.0 0 6 12 18 24 Months NO. ALIVE 201,752 173,906 162,745 152,298 138,077 Figure 4. Adjusted Probability of Survival According to Left Ventricular Ejection Fraction among Patients Who Received or Did Not Receive Beta-Blockers. Patients with ejection fractions under 20 percent had a smaller relative benefit but a larger absolute benefit than patients with normal ejection fractions. The curve for the patients not receiving betablockers who had left ejection fractions of 20 through 49 percent overlaps that for the patients receiving beta-blockers who had left ejection fractions below 20 percent. of an analysis of 5332 Medicare patients without relative contraindications to beta-blockade. 16 Diabetes Mellitus Diabetic patients treated with beta-blockers had a 36 percent reduction in mortality. The present study supports previous work demonstrating that patients with diabetes have a worse outcome after myocardial infarction than those without diabetes. 17,18 However, physicians have been reluctant to treat diabetic patients with beta-blockers because of the risk of worsening the diabetes or of masking hypoglycemic symptoms. Although it is true that blood glucose concentrations may increase with the use of some beta-blockers, 19 our study confirms previous studies that have documented the benefit of beta-blockade in patients with diabetes mellitus. 20,21 Because of the high mortality rate among diabetic patients after myocardial infarction, the absolute survival benefit is particularly large in this group of patients. 494 August 20, 1998

EFFECT OF BETA-BLOCKADE ON MORTALITY IN HIGH-RISK AND LOW-RISK PATIENTS AFTER MYOCARDIAL INFARCTION Pulmonary Disease Patients with a history of chronic obstructive pulmonary disease had a high mortality rate but had a risk reduction of 40 percent with beta-blockade. Even patients with a history of asthma had lower mortality with beta-blockers, although these might be the patients under observation who tolerated the drug. Many physicians withhold beta-blockers from patients with pulmonary disease because of concern about bronchoconstriction. It is true that betablockers cause bronchoconstriction in patients with asthma, 22 and bronchial hyperresponsiveness has even been reported in patients with chronic obstructive pulmonary disease. 23 Indeed, such patients have generally not been evaluated in studies of betablockers because of concern about bronchial hyperresponsiveness. However, the stimuli to hyperresponsiveness in patients with chronic obstructive pulmonary disease may be different from those in patients with asthma, and beta-blockade may not cause bronchoconstriction in patients with chronic obstructive pulmonary disease. 23 This may be the reason that in a study of 50 patients with chronic obstructive pulmonary disease who were given esmolol, there was no change in pulmonary function. 24 It appears clear that beta-blockers should be tried in patients with chronic obstructive pulmonary disease who have had myocardial infarction. It may be necessary in a very few patients to discontinue the drug because of bronchoconstriction, but the potential benefit appears large enough to warrant this small risk. We did not compare b 1 -selective and nonselective drugs. It is possible that they may have different effects on mortality because b 1 -selective drugs have less pulmonary effect. 25 Congestive Heart Failure Whether beta-blockers should be given to patients with both congestive heart failure and myocardial infarction remains controversial. Despite initial concern that the negative inotropic effects of betablockade would be detrimental in patients with heart failure, recent studies have suggested that the longterm effects of beta-blockade are beneficial. 26-28 However, patients with ischemic cardiomyopathy may not fare as well, 29 and many studies of heart failure have evaluated patients with nonischemic cardiomyopathy. There are no randomized studies evaluating the effects of beta-blockers in patients with recent myocardial infarction and poor cardiac function. In one retrospective analysis, propranolol was effective in patients with a history of congestive heart failure, 30 but the sickest patients have traditionally been excluded from studies. The present analysis suggests that patients with the poorest cardiac function benefit as much as patients with better function. Patients with congestive heart failure had a 40 percent reduction in mortality when they received beta-blockers. The relative risk of death decreased by 40 percent in patients with an ejection fraction between 20 percent and 49 percent and by 32 percent in patients with an ejection fraction below 20 percent. However, since patients with the lowest ejection fractions have the highest mortality rate, the absolute benefit for these patients is actually greater than for those with higher ejection fractions. The estimates of ejection fraction are obtained by various methods and should not be considered precise. They do show, however, that patients can benefit from beta-blockade regardless of their ejection fraction. Low-Risk Patients The risks of a drug could conceivably be greater than the benefits in relatively healthy patients. Indeed, the recommendations of the American Heart Association for the treatment of myocardial infarction state that there is continued debate about whether low-risk subjects...should be treated with b-adrenoceptor blockers because their longterm prognosis is extremely favorable irrespective of such therapy. 12 In the present study, however, the lowest-risk subgroups had a decrease in mortality of 40 percent with beta-blockade. These subgroups included patients with normal ejection fractions, patients with low serum creatinine concentrations, and those who had undergone revascularization or thrombolysis. Randomized studies of beta-blocker therapy were performed in the prethrombolytic era, raising the question of whether beta-blockers are beneficial in patients who have undergone reperfusion. 31 Previous smaller analyses of patients receiving thrombolytic agents have not clearly shown a long-term effect on mortality. For example, immediate use of beta-blockers did not improve long-term survival in the II-B substudy of the Thrombolysis in Myocardial Infarction trial. 32 We found a decrease in mortality of 40 percent with the use of beta-blockers in this low-risk subgroup. Because of their lower mortality rates, the absolute benefit is not as great in members of low-risk subgroups, but beta-blockers substantially improved survival in these patients. The question posed by the American College of Cardiology and American Heart Association guidelines for the treatment of patients with acute myocardial infarction appears to have been answered, at least with respect to the predominantly elderly patients analyzed in the present study. Race Black patients with hypertension do not respond as well as whites to treatment with beta-adrenergic blockers. 33,34 The reasons are unknown, but they may be related to differences in beta-adrenoceptor sensitivity. 35 This difference has aroused concern Volume 339 Number 8 495

that blacks may not benefit from beta-blockade after myocardial infarction. In contrast to the b-blocker Heart Attack Trial, which found similar benefits in whites and blacks, 36 the benefit in blacks appeared to be lower in the present study. This could be due to differences in the sympathetic nervous system, differences in the use of the drugs, or other factors. Nevertheless, it should be noted that both blacks and whites had decreased mortality rates when given beta-blockers. Limitations Our study was not a randomized, controlled trial, and its conclusions must be viewed with important reservations. We could not control for characteristics that may influence a physician s decision to prescribe beta-blockers or the patient s decision to comply with a physician s choice of therapy. It is possible that patients who received beta-blockers were healthier than patients who did not or that they differed in some other way that caused them to have lower mortality. They could have had better physicians or have received different medications after discharge. Nevertheless, we attempted to control for or study the interactive effects of most of the known risk factors. It is also possible that there were a substantial number of crossovers between the groups. The data reflect beta-blocker use at hospital discharge; some patients may have started treatment later, and some may have discontinued it. Such an effect, however, would be expected to decrease the power of the study. The consistency among groups suggests that the results of the present analysis reflect the benefit of the intervention. The present analysis also confirms previous studies, but it extends the evidence of the benefit of beta-blockade to groups not previously evaluated. In contrast to other studies, however, our analysis of a large number of patients did not exclude patients at highest risk. This analysis strongly indicates that beta-blockade is an underused therapy for patients who have had a myocardial infarction. Despite a decrease in mortality of between 28 and 40 percent, patients in many subgroups are not being prescribed these drugs. In addition to otherwise healthy patients, those with nontransmural infarction, heart failure, pulmonary disease, and older age are likely to benefit when given beta-blockers after a myocardial infarction. Supported by a grant from the Delmarva Foundation for Medical Care and by the Health Care Financing Administration. We are indebted to Thomas Marciniak, M.D., and William Oetgen, M.D., for their help and support. REFERENCES 1. b-blocker Heart Attack Study Group. The b-blocker Heart Attack Trial. JAMA 1981;246:2073-4. 2. The Norwegian Multicenter Study Group. Timolol-induced reduction in mortality and reinfarction in patients surviving acute myocardial infarction. N Engl J Med 1981;304:801-7. 3. Hjalmarson A, Elmfeldt D, Herlitz J, et al. Effect on mortality of metoprolol in acute myocardial infarction: a double-blind randomised trial. Lancet 1981;2:823-7. 4. Ellerbeck EF, Jencks SF, Radford MJ, et al. Quality of care for Medicare patients with acute myocardial infarction: a four-state pilot study from the Cooperative Cardiovascular Project. JAMA 1995;273:1509-14. 5. Brand DA, Newcomer LN, Freiburger A, Tian H. Cardiologists practices compared with practice guidelines: use of beta-blockade after acute myocardial infarction. J Am Coll Cardiol 1995;26:1432-6. 6. Viskin S, Kitzis I, Lev E, et al. Treatment with beta-adrenergic blocking agents after myocardial infarction: from randomized trials to clinical practice. J Am Coll Cardiol 1995;25:1327-32. 7. Marciniak TA, Ellerbeck EF, Radford MJ, et al. Improving the quality of care for Medicare patients with acute myocardial infarction: results from the Cooperative Cardiovascular Project. JAMA 1998;279:1351-7. 8. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med 1985;13:818-29. 9. Killip T III, Kimball JT. Treatment of myocardial infarction in a coronary care unit: a two year experience with 250 patients. Am J Cardiol 1967; 20:457-64. 10. Yusuf S, Wittes J, Friedman L. Overview of results of randomized clinical trials in heart disease. I. Treatments following myocardial infarction. JAMA 1988;260:2088-93. 11. b-blocker Heart Attack Study Group. A randomized trial of propranolol in patients with acute myocardial infarction. I. Mortality results. JAMA 1982;247:1707-14. 12. Committee on Management of Acute Myocardial Infarction. ACC/AHA Guidelines for the management of patients with acute myocardial infarction. J Am Coll Cardiol 1996;28:1328-428. 13. Anderson JL. Medical therapy for elderly patients who have had myocardial infarction: too little to the late in life? Ann Intern Med 1996;124:335-8. 14. Aronow WS. Prevalence of use of beta blockers and of calcium channel blockers in older patients with prior myocardial infarction at the time of admission to a nursing home. J Am Geriatr Soc 1996;44:1075-7. 15. Gundersen T, Abrahamsen AM, Kjekshus J, Ronnevik PK. Timololrelated reduction in mortality and reinfarction in patients ages 65-75 years surviving acute myocardial infarction. Circulation 1982;66:1179-84. 16. Soumerai SB, McLaughlin TJ, Spiegelman D, Hertzmark E, Thibault G, Goldman L. Adverse outcomes of underuse of b-blockers in elderly survivors of acute myocardial infarction. JAMA 1997;277:115-21. 17. Zuanetti G, Latini R, Maggioni AP, Santoro L, Franzosi MG. Influence of diabetes on mortality in acute myocardial infarction: data from the GISSI-2 study. J Am Coll Cardiol 1993;22:1788-94. 18. Kjekshus J, Gilpin E, Cali G, Blackey AR, Henning H, Ross J Jr. Diabetic patients and beta-blockers after acute myocardial infarction. Eur Heart J 1990;11:43-50. 19. Giugliano D, Acampora R, Marfella R, et al. Metabolic and cardiovascular effects of carvedilol and atenolol in non-insulin-dependent diabetes mellitus and hypertension: a randomized, controlled trial. Ann Intern Med 1997;126:955-9. 20. Gundersen T, Kjekshus J. Timolol treatment after myocardial infarction in diabetic patients. Diabetes Care 1983;6:285-90. 21. Malmberg K, Herlitz J, Hjalmarson A, Ryden L. Effects of metoprolol on mortality and late infarction in diabetics with suspected acute myocardial infarction: retrospective data from two large studies. Eur Heart J 1989; 10:423-8. 22. Tattersfield AE. Respiratory function in the elderly and the effects of beta blockade. Cardiovasc Drugs Ther 1991;4:Suppl 6:1229-32. 23. Woolcock AJ, Anderson SD, Peat JK, et al. Characteristics of bronchial hyperresponsiveness in chronic obstructive pulmonary disease and in asthma. Am Rev Respir Dis 1991;143:1438-43. 24. Gold MR, Dec GW, Cocca-Spofford D, Thompson BT. Esmolol and ventilatory function in cardiac patients with COPD. Chest 1991;100:1215-8. 25. Fogari R, Zoppi A, Tettamanti F, Poletti L, Rizzardi G, Fiocchi G. Comparative effects of celiprolol, propranolol, oxprenolol, and atenolol on respiratory function in hypertensive patients with chronic obstructive lung disease. Cardiovasc Drugs Ther 1990;4:1145-9. 26. Packer M, Bristow MR, Cohn JN, et al. The effect of carvedilol on morbidity and mortality in patients with chronic heart failure. N Engl J Med 1996;334:1349-55. 27. Waagstein F, Bristow MR, Swedberg K, et al. Beneficial effects of metoprolol in idiopathic dilated cardiomyopathy. Lancet 1993;342:1441-6. 28. CIBIS Investigators and Committees. A randomized trial of b-blockade in heart failure: the Cardiac Insufficiency Bisoprolol Study (CIBIS). Circulation 1994;90:1765-73. 29. Woodley SL, Gilbert EM, Anderson JL, et al. b-blockade with bucindolol in heart failure caused by ischemic versus idiopathic dilated cardiomyopathy. Circulation 1991;84:2426-41. 496 August 20, 1998

EFFECT OF BETA-BLOCKADE ON MORTALITY IN HIGH-RISK AND LOW-RISK PATIENTS AFTER MYOCARDIAL INFARCTION 30. Chadda K, Goldstein S, Byington R, Curb JD. Effect of propranolol after acute myocardial infarction in patients with congestive heart failure. Circulation 1986;73:503-10. 31. Becker RC. Beta-adrenergic blockade following thrombolytic therapy: is it helpful or harmful? Clin Cardiol 1994;17:171-4. 32. Roberts R, Rogers WJ, Mueller HS, et al. Immediate versus deferred beta-blockade following thrombolytic therapy in patients with acute myocardial infarction: results of the Thrombolysis in Myocardial Infarction (TIMI) II-B Study. Circulation 1991;83:422-37. 33. Materson BJ, Reda DJ, Cushman WC, et al. Single-drug therapy for hypertension in men: a comparison of six antihypertensive agents with placebo. N Engl J Med 1993;328:914-21. [Erratum, N Engl J Med 1994;330:1689.] 34. Prisant LM, Mensah GA. Use of beta-adrenergic receptor blockers in blacks. J Clin Pharmacol 1996;36:867-73. 35. Johnson JA, Burlew BS, Stiles RN. Racial differences in beta-adrenoceptor-mediated responsiveness. J Cardiovasc Pharmacol 1995;25:90-6. 36. Haywood LJ. Coronary heart disease mortality/morbidity and risk in blacks. I. Clinical manifestations and diagnostic criteria: the experience with the Beta Blocker Heart Attack Trial. Am Heart J 1984;108:787-93. Volume 339 Number 8 497